(commentary)
Things are starting to settle in and details are getting fleshed out on Nikon’s unexpected announcement of twins yesterday. I’ll keep updating the data pages for the two products as I learn more, so consider them a work in progress for the time being.
Some random thoughts on what Nikon has given us:
- The 4K video folk are all grumbling. Why? Because both cameras use crops to achieve 4K video (D5 is 1.5x, D500 is 2.2x). There’s no perfect solution to 4K other than to have a perfectly sized sensor for 4K, which neither camera has. You really need 7680 x 4320 pixels (33mp) to have a full Bayer pattern for every recorded pixel. So what you have to do when you don’t have that is come up with some sort of scaling or subsampling capability, and none of the choices to do so are perfect. It’ll take close examination of the new cameras’ output to determine how well this works, but the added crop factor while shooting video really messes up the lens selection/choice, especially for the D500. Nikon’s going to wish they had made a lot of wide angle DX lens options ;~).
- The 4K video folk are all grumbling. Why? Wait, didn’t I just write that? No, they’re grumbling about multiple things. The other big one is the short record times (3 minutes) for 4K video on the D5 (on the D500 camera apparently splits files up, but can record 4K longer). It doesn’t seem to be a heat issue, as you can output to the HDMI port indefinitely, so it probably has to do with the video compression and file size limitations. As much as people wanted Nikon to succeed at “video,” Nikon is still executing at the margins.
- While we’re on video: the cameras appear to slightly overscan 1080P, and have the ability to remove left/right, up/down, and rotation movement at the sensor plane via electronic recalculation of the frame area, something called “electronic VR.” Update: an earlier version of this suggested this works for 4K. Nope, only the regular HD movie modes.
- A number of questions still linger, and NikonUSA doesn’t yet seem to be able to answer them. Using XQD the buffer should be 200 shots if you’re using any form of compressed NEF. Moving to uncompressed drops you to 79 shots. Of course, if you’re using any of the secondary slot functions, your buffer is likely limited to that of the slower card, which is where some confusion is occurring. No one seems to know how compatible the new wireless flash is to older cameras. I’d assume that all i-TTL cameras should work with it via radio, but those older cameras also don’t have the menu capabilities to drive the new flash using camera settings, so there’s a gray area that’s as yet unresolved.
- Sure enough, my In Box is starting to fill up with “I want one, but I’m going to wait until the early adopters find all the quality control problems” messages. If Nikon doesn’t get as many initial orders for the cameras as they expect, it’s their own fault. The Nikon enthusiast seems very impressed by the two new models and the performance they suggest, but have burned one too many times with having to send products back to Nikon for fixes of things that should have been caught before shipping. I’ve written about this many times before: Nikon is their own worst enemy: they constantly create frictions that reduce their actual results. Whether those frictions represent 1%, 5%, 10%, or more of potential sales doesn’t matter. What matters is that at a time when the camera industry is under pressure, Nikon is operating sub optimally to how they could have.
- Is the area covered by the autofocus sensor areas larger in the D5 than the D4 (and other FX DSLRs)? Yes. Nikon claims something near 30% larger in overall area, though this mostly extends to the sides of the long axis. On the D500 this makes the coverage as close to edge to edge for the long axis as we’ve ever seen in a DSLR. Why isn’t the area even bigger? Frankly, I’m surprised that Nikon was able to squeeze out as much additional area as they did. The geometry of how phase detect works makes it difficult to cover other than a central area. Nikon had been pursuing a couple of different methods of expanding the area. Amazingly, deeply finessing the old approach worked well enough for them to select it.
- Button illumination lives. It’s on both models (first appeared on the D4). Now if we could only program colors and flashing patterns ;~).
- I wouldn’t get too excited about those sky high ISO values. Let me try to explain. Current digital sensors max out with electron wells of about 100k e-. Even if we assume state of the art electronics, I’ll bet that we have residual and read problems at about the 5-10 e- level. And, of course, photons are random, so we have quantum shot noise, too. If you’re having to shoot at 1m+ ISO, how many photons are we talking about potentially capturing in a short exposure? 10? We very well could be at a SN ratio of 1:1. We’re certainly going to be nowhere close to the SN ratio of 80:1+ I really want in my data. So what I’ll be looking at is situations where there’s light that is more likely to generate enough photons to be collected. In particular, I’ll be looking closely at the ISO 6400 to ISO 12,800 performance of these cameras. What I hope to see is that I’m quantum shot noise limited, not sensor limited, when I shoot at those ISO values. That would actually be a big deal with a DX body, though not so much with a full frame one.
- My initial visual impression is that the D500 appears more rushed. The sample images struck me to be what looked to me as early JPEG engine rendering (e.g. early firmware). My guess is that the sensor for this camera matured after the D5 sensor and Nikon is still working towards tuning it.
- I’m not sure why Nikon pre-announced the Keystone 360. All this did is give competitors to see what Nikon is up to in this space. Basically, Nikon gave GoPro a several month window to prepare their marketing response.
Support this site: If you’d like one B&H link to all the Nikon products announced at CES, click here.
